Electron beam induced current (EBIC) and atom probe tomography (APT) were used in this study to determine electrical activities and impurity compositions at extended defects in multicrystalline silicon (mc-Si) samples. The results provide, for the first time, information regarding the chemical species present at defects whose electrical activity has previously been measured. A new APT specimen fabrication process was developed with the ability to select a specific defect for APT analysis. Development of the APT specimen fabrication process proceeded by first selecting and optimising the preferential etching for nano-scale defect delineation. Three etchants were evaluated, namely Secco, Sirtl and Dash, from which the Secco etch was selected. Three parameters were optimised to produce etch pits with geometries that meet the requirements imposed by APT specimen fabrication methods. The optimum parameters were 0.05M potassium dichromate concentration, 20°C etch temperature, and 30sec etch time. In the second stage, marking techniques were developed in order for the defects to be located throughout the APT specimen fabrication process. However, it became apparent that the conventional APT specimen fabrication method could not be used to fabricate APT specimens containing selected defects in a mc-Si sample. This led to the development of a novel APT specimen fabrication approach which allowed APT specimens to be fabricated, reproducibly, containing grain boundaries and isolated dislocations. In order to evaluate accurately iron contamination in mc-Si, four atom probe parameters were optimised to maximise detection sensitivity: the evaporation rate, the laser beam energy, the pulse repetition rate and the specimen temperature. The optimisation process can be divided in to two parts. In the first part, a matrix of pre-sharpened single-crystal silicon specimens was subjected to a variety of experimental parameters. The optimised parameters were determined to be 0.3% evaporation rate, 0.5nJ beam energy, 160kHz repetition rate and 55K specimen temperature. The second part was to determine the iron detection efficiency –the percentage of detected Fe ions that can be correctly identified as Fe– and sensitivity using these parameters to analyse a specially prepared iron calibration specimen. The values were determined to be a detection efficiency of about 35% and sensitivity of 54ppm or 2.70x10¹⁸ atom/cm³. The APT specimen fabrication process and the optimised APT analysis parameters were used to analyse four extended defects in mc-Si samples subjected to three different processing conditions, namely gold-contaminated, as-grown and phosphorus diffusion gettering (PDG). The important aspects of the analysis are listed below: • Gold was not detected at the grain boundary and its associated dislocations in the gold-contaminated specimen. The binding enthalpy of gold to such defects is thus less than 0.63eV. • Iron was not detected in any specimen. • Copper was observed at the grain boundary in the as-grown specimen in the form of individual atoms as well as clusters with diameters ranging between 4nm and 9nm. The electrical activity of the grain boundary was about 58%. • Nickel and carbon were detected at the grain boundary in the post-PDG specimen with the former having platelet structures with diameters and thicknesses ranging between 4nm-7nm and 2nm-4nm, respectively. The recombination strength of the defect was about 22%. • Two nickel clusters were found at the isolated dislocation in the post-PDG specimen. The clusters were spherical with an average diameter of 10nm. The distance between the two clusters was 35nm. The recombination strength of the defect was about 4%.

A new materials system of a single layered Co(W) barrier/liner coupled with a Cu(Mn) alloy seed was investigated. Atom probe tomography visualized the sub-nanoscale structure of Cu(Mn)/Co(W) system, and thereby revealed Cu diffusion behavior of Co(W). Grain boundaries of Co were found to be the diffusion path, and successfully stuffed by W. Mn in Cu(Mn) also segregated to stuff the grain boundaries of Co. Combination of these two additives enabled high barrier property against Cu diffusion of Cu(Mn)/Co(W). Foreseeing tiny and high-aspect-ratio Cu interconnect features, Cu(Mn)/Co(W) was fabricated by ALD/CVD processes. To maximize the performance, minor impurities of the film incorporated from the ligand of the precursors were controlled by precursor selection. Thin, conformal, and smooth films were finally demonstrated onto a trench substrate.

While amorphous phases have been reported in immiscible alloy systems, there is still some controversy regarding the reason for the stabilization of these unusual amorphous phases. Direct evidence of nanoscale phase separation within the amorphous phase forming in immiscible Cu-Nb alloy thin films using 3D atom probe tomography has been presented. This evidence clearly indicates that the nanoscale phase separation is responsible for the stabilization of the amorphous phase in such immiscible systems since it substantially reduces the free energy of the undercooled liquid (or amorphous) phase, below that of the competing supersaturated crystalline phases. The devitrification of the immiscible Cu-Nb thin film of composition Cu-45% Nb has been studied in detail with the discussion on the mechanism of phase transformation. The initial phase separation in the amorphous condition seems to play a vital role in the crystallization of the thin film. Detailed analysis has been done using X-ray diffraction, transmission electron microscopy and 3D atom probe tomography.

This work was produced as part of a multidisciplinary study of the corrosion of zirconium alloys undertaken by a consortium of universities working in the MUZIC program; Oxford, Manchester and The Open University. The objective of the project as a whole was to further the understanding of the mechanisms of the breakaway oxidation process and to characterise these corrosion processes within a number of fuel rod cladding materials. This thesis describes laser 3D atom probe characterisation of the nano-scale chemical redistribution of oxygen and other solutes that occurs at the metal-oxide interface during corrosion, and a large body of technique development that was required to achieve this goal. The development of the metal-oxide interface of ZIRLO, a Zr-Nb-Sn-Fe-O alloy, is followed by generating 3D atomic scale reconstructions at four different stages of corrosion. The formation of a sub-oxide ZrO layer is seen during pre-transition oxide development. The ZrO interfacial layer is consumed by the rapid formation of oxide after the breakaway transition. After transition the chemistry of the interface is similar to the early pre-transition case, although an oxygen-saturated layer of metal adjacent to the interface formed during corrosion remains. The ZrO interfacial layer (Zr-ZrO-ZrO₂) and the region of oxygen-saturated material ahead of the metal-oxide interface alter the distribution of minor alloying additions such as niobium and iron. The ZrO layer increases the acceptance of niobium into the oxide, which is otherwise seen to be rejected at the Zr-ZrO2 interface along with iron. Niobium is seen to precipitate out of solution as nano-scale particles near the interface after around 100 days of corrosion. This is not seen in the bulk metal matrix of the corroded material due to the absence of other factors driving the process: the stress at the interface and a very high oxygen concentration in the metal ahead of the interface. The nano-scale niobium particles are found to be of a meta-stable composition. Iron is seen to redistribute in the corroded material and can be correlated with the local oxygen concentration. Similarities are seen in the behaviour of solutes within pre-transition ZIRLO and Zircaloy-4 (Zr-Sn-Fe-O). In both cases no redistribution of tin is seen at the metal-oxide interface. A Zr-Nb-Ti alloy with very poor corrosion resistance was also analysed in this way, and the similarities and differences with chemically-similar ZIRLO are discussed. The segregation of solutes to grain boundaries and solute clustering within the matrix are also examined before and after corrosion.

The increasing technological need to push service conditions of structural materials to higher temperatures has motivated the development of several alloy systems. Among them, superalloys are an excellent candidate for high temperature applications because of their ability to form coherent ordered precipitates, which enable the retention of high strength close to their melting temperature. The accelerated kinetics of solute diffusion, with or without an added component of mechanical stress, leads to coarsening of the precipitates, and results in microstructural degradation, limiting the durability of the materials. Hence, the coarsening of precipitates has been a classical research problem for these alloys in service. The prolonged hunt for an alternative of nickel base superalloys with superior traits has gained hope after the recent discovery of Co-Al-W based alloys, which readily form high temperature g precipitates, similar to Ni base superalloys. In the present study, coarsening behavior of g precipitates in Co-10Al-10W (at. %) has been carried out at 800°C and 900°C. This study has, for the first time, obtained critical coarsening parameters in cobalt-base alloys. Apart from this, it has incorporated atomic scale compositional information across the g/g interfaces into classical Cahn-Hilliard model for a better model of coarsening kinetics. The coarsening study of g precipitates in Ni-14Al-7 Cr (at. %) has shown the importance of temporal evolution of the compositional width of the g/g interfaces to the coarsening kinetics of g precipitates. This study has introduced a novel, reproducible characterization method of crystallographic study of ordered phase by coupling of orientation microscopy with atom probe tomography (APT). Along with the detailed analysis of field evaporation behaviors of Ni and Co superalloys in APT, the present study determines the site occupancy of various solutes within ordered g precipitates in both Ni and Co superalloys. This study has explained the role of structural and compositional gradients across the precipitates (g)/matrix (g) interfaces on the coarsening behavior of coherent precipitates in both Ni and Co-base superalloys. The observation of two interfacial widths, one corresponding to a structural order-disorder transition, and the other to the compositional transition across the interface, raises fundamental questions regarding the definition of the interfacial width in such systems. The comparative interface analysis in Co and Ni superalloy shows significant differences, which gives insights to the coarsening behaviors of g precipitates in these alloys. Hence, the principal goal of this work is to compare and contrast the Co and Ni superalloys and also, to accommodate atomic scale information related to transitions across interfaces to coarsening models for a better practical applicability of coarsening laws to various alloys.

Titanium nitride (TiN) films are widely applied as diffusion barrier layers in microelectronic devices. The continued miniaturization of such devices not only poses new challenges to material systems design, but also puts high demands on characterization techniques. To gain understanding of diffusion processes that can eventually lead to failure of the barrier layer and thus of the whole device, it is essential to develop routines to chemically and structurally investigate these layers down to the atomic scale. In the present study, model TiN diffusion barriers with a Cu overlayer acting as the diffusion source were grown by reactive magnetron sputtering on MgO(001) and thermally oxidized Si(001) substrates. Cross-sectional transmission electron microscopy (XTEM) of the pristine samples revealed epitaxial, single-crystalline growth of TiN on MgO(001), while the polycrystalline TiN grown on Si(001) exhibited a [001]-oriented columnar microstructure. Various annealing treatments were carried out to induce diffusion of Cu into the TiN layer. Subsequently, XTEM images were recorded with a high-angle annular dark field detector, which provides strong elemental contrast, to illuminate the correlation between the structure and the barrier efficiency of the single- and polycrystalline TiN layers. Particular regions of interest were investigated more closely by energy dispersive X-ray (EDX) mapping. These investigations are completed by atom probe tomography (APT) studies, which provide a three-dimensional insight into the elemental distribution at the near-interface region with atomic chemical resolution and high sensitivity. In case of the single-crystalline barrier, a uniform Cu-enriched diffusion layer of 12 nm could be detected at the interface after an annealing treatment at 1000 °C for 12 h. This excellent barrier performance can be attributed to the lack of fast diffusion paths such as grain boundaries. Moreover, density-functional theory calculations predict a stoichiometry-dependent atomic diffusion mechanism of Cu in bulk TiN, with Cu diffusing on the N-sublattice for the experimental N/Ti ratio. In comparison, the polycrystalline TiN layers exhibited grain boundaries reaching from the Cu-TiN interface to the substrate, thus providing direct diffusion paths for Cu. However, the microstructure of these columnar layers was still dense without open porosity or voids, so that the onset of grain boundary diffusion could only be found after annealing at 900 °C for 1 h. The present study shows how to combine two high resolution state-of-the-art methods, TEM and APT, to characterize model TiN diffusion barriers. It is shown how to correlate the microstructure with the performance of the barrier layer by two-dimensional EDX mapping and three-dimensional APT. Highly effective Cu-diffusion barrier function is thus demonstrated for single-crystal TiN(001) (up to 1000 °C) and dense polycrystalline TiN (900 °C).

Doctor of Philosophy (PhD) / The evolution of atomistic-level nanostructure during the early stages of both standard, high-temperature T6 heat treatment, and low-temperature secondary ageing after interruption of the former (T6I4), has been investigated in rapid hardening Al-Cu-Mg alloys using a variety of microscopy and microanalytical techniques, including transmission electron microscopy (TEM), positron annihilation spectroscopy (PAS) and atom probe tomography (APT). In order to carry out this objective, quantitative data-analysis methods were developed with respect to new cluster-finding algorithms, specifically designed for use with three-dimensional APT data. Prior to this detailed characterisation work, the actual thermal impact from both heat treatment and quenching of small, lab-scale specimens was determined through correlation of both experimental results and calculations that modelled the heat transfer conditions using the lumped capacitance method. Subsequently, the maximum diffusion distance by random walk of the solute atoms was calculated for these periods, bearing significance on the propensity for these atoms to have the ability to cluster together, rather than segregate to the dislocation loops in the microstructure, which have a relatively larger interspacing distance. Age-hardening curves for the Al-1.1Cu-xMg (x = 0, 0.2, 0.5, 0.75, 1.0, 1.7 at.%) alloys at 150ºC show that the rapid hardening phenomenon (RHP) exists for Mg compositions ≥ 0.5Mg. Given that zone-like precipitate structures were unable to be detected by TEM or APT during the early stages of ageing at 150ºC, and that statistically significant dispersions of clusters were found in the APT data after ageing for 60 s, the RHP is attributed to these clustering reactions. Identification of clusters in the APT data has been achieved using the core-linkage algorithm and they have been found to be quite small, containing only a few atoms up to a couple of tens of atoms. The RHP is governed by some critical number density of both Mg clusters and Cu-Mg co-clusters of a critical size, whereas Cu clusters do not contribute significantly to the hardening mechanism. Significance testing indicates that Mg clusters are more significant at smaller clusters sizes and Cu-Mg co-clusters more important at larger cluster sizes. Hardness results also confirm the existence of rapid early hardening during secondary ageing at 65ºC in Al-1.1Cu-1.7Mg. The mechanism of secondary rapid hardening involves a combination of both secondary clustering from solute (mainly Mg atoms) residual in solution, and pre-existing amorphous primary clusters that have slower growth kinetics at the lower secondary ageing temperature. The latter occurs mainly by vacancy-assisted diffusion of Mg atoms as evidenced by the gradual increase of the Mg:Cu ratio of co-clusters. From an alloy design point of view it is important to fully understand the solute distribution in the microstructure to be able to subsequently optimise the configuration for enhanced material properties. The change in dispersion of solute atoms during ageing was determined by combining calculations of % vacancy-solute associations with detailed measurements of the dislocation loops to estimate the solute distribution within the microstructure. The implication of the balance of solute atoms segregated to the loops compared with that in the matrix is then discussed in the context of hardnening mechanisms.

In microelectronics, the device size continues to shrink to improve the performance and functionality, which sets technical challenges for the integrated circuit (IC) fabrication. Novel materials and processing techniques are developed to maintain excellent device performances and structural reliability. Cobalt-based thin films possess numerous applications in microelectronics with the potential to enhance the device performance and reliability. This thesis explores the fabrication, characterization and application of cobalt-based thin films for microelectronics. Chemical vapor deposition (CVD) technique has been applied for depositing cobalt-based thin films, because CVD can produce high quality thin films with excellent conformality in complex 3D architectures required for future microelectronics. / Engineering and Applied Sciences

Doctor of Philosophy (PhD) / The evolution of atomistic-level nanostructure during the early stages of both standard, high-temperature T6 heat treatment, and low-temperature secondary ageing after interruption of the former (T6I4), has been investigated in rapid hardening Al-Cu-Mg alloys using a variety of microscopy and microanalytical techniques, including transmission electron microscopy (TEM), positron annihilation spectroscopy (PAS) and atom probe tomography (APT). In order to carry out this objective, quantitative data-analysis methods were developed with respect to new cluster-finding algorithms, specifically designed for use with three-dimensional APT data. Prior to this detailed characterisation work, the actual thermal impact from both heat treatment and quenching of small, lab-scale specimens was determined through correlation of both experimental results and calculations that modelled the heat transfer conditions using the lumped capacitance method. Subsequently, the maximum diffusion distance by random walk of the solute atoms was calculated for these periods, bearing significance on the propensity for these atoms to have the ability to cluster together, rather than segregate to the dislocation loops in the microstructure, which have a relatively larger interspacing distance. Age-hardening curves for the Al-1.1Cu-xMg (x = 0, 0.2, 0.5, 0.75, 1.0, 1.7 at.%) alloys at 150ºC show that the rapid hardening phenomenon (RHP) exists for Mg compositions ≥ 0.5Mg. Given that zone-like precipitate structures were unable to be detected by TEM or APT during the early stages of ageing at 150ºC, and that statistically significant dispersions of clusters were found in the APT data after ageing for 60 s, the RHP is attributed to these clustering reactions. Identification of clusters in the APT data has been achieved using the core-linkage algorithm and they have been found to be quite small, containing only a few atoms up to a couple of tens of atoms. The RHP is governed by some critical number density of both Mg clusters and Cu-Mg co-clusters of a critical size, whereas Cu clusters do not contribute significantly to the hardening mechanism. Significance testing indicates that Mg clusters are more significant at smaller clusters sizes and Cu-Mg co-clusters more important at larger cluster sizes. Hardness results also confirm the existence of rapid early hardening during secondary ageing at 65ºC in Al-1.1Cu-1.7Mg. The mechanism of secondary rapid hardening involves a combination of both secondary clustering from solute (mainly Mg atoms) residual in solution, and pre-existing amorphous primary clusters that have slower growth kinetics at the lower secondary ageing temperature. The latter occurs mainly by vacancy-assisted diffusion of Mg atoms as evidenced by the gradual increase of the Mg:Cu ratio of co-clusters. From an alloy design point of view it is important to fully understand the solute distribution in the microstructure to be able to subsequently optimise the configuration for enhanced material properties. The change in dispersion of solute atoms during ageing was determined by combining calculations of % vacancy-solute associations with detailed measurements of the dislocation loops to estimate the solute distribution within the microstructure. The implication of the balance of solute atoms segregated to the loops compared with that in the matrix is then discussed in the context of hardnening mechanisms.

Les alliages Al-Cu-Li sont particulièrement attractifs pour les applications aéronautiques du fait de leur faible densité, haute limite d'élasticité et bonne ténacité. Ils reçoivent une attention particulièrement importante actuellement, depuis le développement de la troisième génération qui contient des concentrations relativement élevées pour le cuivre et relativement basses pour le Li. Ces nouveaux alliages sont caractérisés par une dureté élevée, une bonne résistance à la fatigue et une bonne stabilité thermique. La phase principale de durcissement est la phase T1 – Al2CuLi qui se présente sous la forme de plaquettes d'environ 1 nm d'épaisseur et 50 nm de diamètre, situées sur les plans {111} de la matrice avec une structure hexagonale. La germination efficace de cette phase durcissante entre en compétition avec d'autres précipités des sous-systèmes constituant ces alliages (comme Al-Cu et Al-Li), et nécessite des conditions particulières, en particulier la présence de dislocations (introduites par pré-déformation) et d'éléments d'alliage mineurs (Mg, Ag, Zn). Bien qu'il soit connu depuis longtemps que l'addition de ces éléments favorise la cinétique de précipitation dans ces alliages et le durcissement associé, leurs mécanismes d'action sont encore très mal compris.Dans ce contexte, l'objectif de la thèse est d'évaluer systématiquement l'effet des additions mineures de Mg, Ag et Zn sur la germination, la cinétique de précipitation et le durcissement correspondant. La caractérisation détaillée de la microstructure est utilisée pour comprendre les mécanismes de modification de la microstructure par les éléments mineurs. Les mesures de la diffusion des rayons X à petits angles et la DSC fournissent respectivement la cinétique de précipitation et la séquence de formation des phases. La microscopie électronique en transmission, utilisée en mode conventionnel, en résolution atomique et en mode de cartographie chimique met en évidence la structure et la distribution spatiale des phases. La dureté donne accès au durcissement. Des matériaux à gradient de concentration ont été élaborés et caractérisés pour évaluer l'effet de la concentration des alliages sur la précipitation et le durcissement.Les résultats mettent en évidence que le Mg est l'élément le plus efficace pour accélérer la cinétique de précipitation et de durcissement. L'addition d'Ag et de Zn augmente également la cinétique de précipitation mais dans une moindre mesure. L'addition de Mg change la séquence de précipitation tout au long de la séquence de vieillissement. La différence principale liée à la présence de Mg pour les premiers stades de traitement thermique est observée par rapport à la précipitation sur les dislocations. Dans les alliages qui contiennent du Mg, les dislocations sont décorées par des phases précurseur contenant de Cu et Mg. Par contre dans les alliages sans Mg celles-ci sont associés à des zones GP qui évoluent ensuite en précipités θ'. Cette différence est attribuée à la germination favorable de T1 sur les phases précurseur de Cu/Mg dans les alliages contenant du Mg, et par la saturation des sites de germination hétérogène par θ' dans les alliages sans Mg.L'augmentation de dureté associée à l'addition d'Ag et Zn est attribuée à une fraction volumique plus élevé de la phase T1. Ag est ségrège à l'interface entre T1 et la matrice et Zn est incorporé dans la structure de T1. Ces résultats suggèrent que les additions de Zn et Ag stimulent la formation de T1.L'influence de la concentration en éléments d'addition mineurs a été caractérisée par une approche résolue en temps et en espace, sur les matériaux contenant un gradient en composition. Cela révèle que l'effet de l'addition de Mg sur la précipitation se produit à une valeur seuil de ~0.1% en poids, suggérant que cela est la concentration nécessaire pour germer des phases précurseur sur les dislocations dans les premiers stades de la précipitation. / Al-Cu-Li alloys are very attractive for aerospace applications alloys due to their low density, high modulus and high strength. They are experiencing a strong interest since the so-called 3rd generation alloys, with relatively high Cu and low Li content, have been developed with high toughness, fatigue resistance and thermal stability. The main precipitating phase in these alloys is the T1-phase which precipitates on {111}Al-planes with a hexagonal structure. It is known that obtaining a fine dispersion of T1, and hence a high strength requires the presence of dislocations as nucleation sites. In addition, commercial Al-Cu-Li alloys contain several minor alloying elements such as Mg, Ag and Zn, which help reaching the desired properties. Although the effect of these minor additions on precipitation of T1 has been characterized, it has not been understood yet.In this context the aim of this thesis is to systematically investigate the effect of minor additions of Mg, Ag and Zn on precipitation nucleation, precipitation kinetics and related strengthening, and to use a detailed characterization of the microstructure to understand the mechanisms by which the modifications induced by these minor additions take place. In-situ Small-Angle X-ray Scattering and Differential Scanning Calorimetry provide the precipitation kinetics and sequence, respectively. Transmission Electron Microscopy, both in conventional mode, atomically-resolved and in chemical mapping mode, reveals the structure and distribution of phases. Hardness gives access to the strengthening. Compositionally gradient materials are fabricated and characterized to evaluate the effect of alloy composition on precipitation and strengthening.The results reveal that Mg is most effective in order to enhance precipitation kinetics and hardening. Additional Ag and Zn further enhance precipitation kinetics but to a lower extent. The addition of Mg changes the precipitation sequence at all times of ageing. The main differences in early aging conditions are observed with respect to precipitation on dislocations. In Mg-containing alloys, dislocations are decorated by Cu-Mg precursor phases, whereas dislocations in Mg-free alloys are mainly associated to GP-zones which evolve subsequently into θ'-phase. In fully precipitated conditions the microstructure of Mg-containing alloys is dominated by the T1 phase, whereas that of Mg-free alloys is dominated by the θ'-phase. This difference is attributed to the favourable nucleation of T1 on Mg-Cu precursor phases in the Mg-containing alloys, and to the consumption of T1-heterogeneous nucleation sites by the θ'-phase in the Mg-free alloys.The increase of hardness associated to the addition of Ag and Zn is associated to a higher volume fraction of the T1-phase. Ag was found to segregate at the T1/matrix interface and Zn was incorporated into the T1-phase, so that it is assumed that their additions stimulate the formation of T1.The influence of the concentration of the minor solute additions has been characterised by combined space and time-resolved experiments on compositionally gradient materials. It reveals that the effect of an Mg addition on precipitation occurs at a threshold level of ~0.1wt%, suggesting that this concentration is that necessary to form the precursor phase at the dislocations during early ageing.